Planning of Collision-Free Trajectories for Robotic Manipulation and Management of DOFs for Positional Targets

This work aims to identify a suitable ROS-based motion planning pipeline able to generate optimal collision-free trajectories for robotic manipulation and manage and exploit additional degrees of freedom for positional targets. Given an arbitrary target pose, the manipulator is granted an additional rotational degree of freedom along one of the end effector frame's axes. Suitable performance indices are used to select one of the configurations that the robot may attain to reach the assigned positional target. The performances of several indices and motion planning pipelines are evaluated through benchmarks on a 6-axis anthropomorphic manipulator. The experimental results show that, among the various candidates, the ones that achieved the best performance are the distance-based indices and the SBL motion planner with STOMP as a post-processor. The definitive planning pipeline is validated in real manipulation scenarios, where it manages to fulfill the assigned tasks.